Radiographic analysis and virtual cleaning of a bioarchaeological remain enclosed in mineral deposits from a limestone cave

The Bedara cave is located on Mt. Žumberačka Gora (NW Croatia) [1] (Fig. 1a). The area lies at the transition between the southeastern Alps, the northwestern Dinarides and the Pannonian Basin. Altitudes range from 180 to 1,178 m above sea level, with Sveta Gera (1,178 m above mean sea level) as the highest peak. Due to its predominantly carbonate-rich composition, this area tends to form karst [2]. Consequently, numerous karst relief forms, both on the surface, and underground are known, including 205 registered speleological features (pits and caves) [3]. The Bedara cave was discovered in 2002 by cavers of the Caving Club Samobor. It is one of the largest caves of northwestern Croatia, with 1,593 m of explored passages reaching down to a depth of 113 m. Sediments occur as speleothems, clay sediments, silt, sand, debris, gravel, and stone blocks. The cave features several vertical passages without a connection to the surface, as seen at the discovery site at a depth of 47 m (Fig. 1b, point 4), where there is also a continuous flow of drip water (from a narrowing in Fig. 1b, point 5). Here, a proximal part of a right femur, heavily encrusted in calcite carbonate, was discovered during a cave research trip in 2017 within a large calcite deposit (Fig. 2). Based on its external morphology, it could have been of human origin. Further animal skeletal remains were found in this location, suggesting that at a certain period, a vertical passage, functioning as a natural trap, had existed.

The specimen, a proximal femoral fragment with a transversely broken shaft, is completely covered by a thick layer of calcium carbonate. Whilst the anatomical identification and the orientation of the bone were relatively evident due to overall morphological characteristics (generic shape consisting of a cylindrical shaft and spherical head), the thick surrounding concretion altered its relative size, giving the impression of a larger animal (or human), and hid all finer morphological characteristics. A precise taxonomic identification through standard anatomical analysis was, therefore, impossible without prior cleaning of the bone surface. To avoid aggressive chemical or mechanical removal of the calcium carbonate, likely damaging the bone fragment, we used clinical computed tomography (CT) to scan the specimen, and then virtually cleaned the obtained volumetric dataset. Thereby, we revealed the bone fragment’s original morphology and partially also its original surface texture.

The specimen was brought to the University Hospital Centre “Zagreb” (Zagreb, Croatia) and imaged with a clinical computed tomography (CT) scanner (Light Speed Ultra, General Electric, Chicago, USA) using 8 × 1.25 mm collimation, 0.75 pitch, and 120 kVp. Image contrast was considered insufficient for segmentation in the preview on the workstation, and no reconstructions were performed. A different clinical CT scanner (Sensation 40, Siemens Healthineers, Erlangen, Germany) with 40 × 0.6 mm collimation, 0.9 pitch, and 140 kVp at 390 mAs was used for a second scan, producing substantially better contrast (Fig. 3a). Axial sections were reconstructed with a bone weighted kernel (B70s), a slice thickness of 0.6 mm, and an increment of 0.3 mm. A total of 682 slices was generated for coverage of the entire sample.

Nevertheless, segmentation of the incrusted bone remained a particular challenge due to the sometimes very similar densities of the partially fossilised bone fragment and surrounding concretions (Fig. 3a). Hounsfield units (HU) threshold-based or region-growing-based segmentation algorithms were not practicable. A combined, manual and HU threshold-based segmentation protocol, newly developed for this project by one of the authors (PE), was therefore applied. For later validation, segmentation was independently carried out by three readers (R1–3) using the dedicated open-source medical DICOM image viewing software (Horos, v3.2.1, www.​horosproject.​org): First, in an axially reconstructed series of 702 slices, a free-form region of interest (ROI), most closely corresponding to the demarcation of the bone fragment, was manually drawn on every tenth slice. Particular care was taken not to include components of the calcite crust or omit parts of the bone fragment; furthermore, ROIs were drawn clockwise starting at a twelve o’clock position (Fig. 3b). Second, missing ROIs were interpolated on the remaining slices. After this step, added ROIs were checked by scrolling through the entire data set and manually corrected where necessary. Once each slice contained a ROI circumscribing the bone fragment’s delimitation, the “ROI Volume/Compute Volume...” function allowed a first visual inspection of the virtually cleaned bone fragment. Third, the contents of every ROI on each slice, i.e., the actual bone fragment, was deleted with the “ROI Volume/Erase Content” function, and the resulting data set was saved as a new DICOM series (Fig. 3c). The newly generated and stored template DICOM series, only representing the calcite concrement, was then subtracted from the original (unaltered) axially reconstructed series, resulting in a dataset only showing containing the virtually cleaned femoral fragment (Fig. 3d). Last, using conventional HU threshold segmentation, a three-dimensional (3D) surface rendering was created and exported in the “standard tessellation language” file format for subsequent evaluation and dimensioning with the 3D design software (Rhinoceros 3D, Version 5.4.1, Robert McNeel & Associates, Seattle, WA, USA) (Fig. 3e).

The taxonomic assessment was performed by a zooarchaeologist (SR), based on measurements and renderings of the virtual 3D models resulting from CT image segmentation. Measures were taken following von den Driesch [4], and results were then compared with published metric data for species in question. Anatomical terminology used follows standards referenced in the Nomina Anatomica Veterinaria (5th edition) [5].

A direct radiocarbon date of a small sample of the exposed shaft was conducted in conventional 14C age, and fraction corrected using AMS ¹³C. Calibrated age ranges using Intcal13.14c (1 sigma) [6] were 3637–3627 (probability distribution 0.133), 3591–3527 (probability distribution 0.867), (2 sigma) 3647–3516 (probability distribution 0.976), 3408–3405 (probability distribution 0.003) and 3398–3384 (probability distribution 0.020).

The resulting 3D models of R1–3 were later also statistically compared, by point-to-point comparison of the respective point-clouds, using the dedicated open-source software (CloudCompare V2.6, www.​cloudcompare.​org) to validate the accuracy of the resulting models (Fig. 4a). For statistical comparison of the inter-reader correlation, the number of points of the point-clouds was reduced to 1 million points per model. Descriptive statistical analysis and boxplots were created using the dedicated software (SPSS Statistics, Version 24, IBM, Armonk, NY, USA), after importing comparative statistical from CloudCompare in the “comma-separated values” data format.

The combined segmentation protocol proved to be practicable, and the 3D models created by R1–3 achieved excellent congruence, with over 68% of all points lying within a range of less than 1 mm. Mean deviations ranged from 0.29 mm (R3 versus R1) to 0.68 mm (R2 versus R3). Maximum deviations ranged from 0.8 (R3 versus R1) to 2.59 (R2 versus R3). Most pronounced deviations between all three models were observed at the epiphyseal level; however, without affecting structures of osteometric relevance (Fig. 4a). Considering that the original CT scan had an anisotropic voxel size of 0.33 × 0.33 × 0.67 mm (x, y, and z, respectively), mean deviations even remain within the limitations of the CT scan’s resolution. In brief, the resulting model provided sufficient data and visualisation for the morphometric and taxonomic study of the specimen (Fig. 3e).

The resulting images show that the bone fragment is well preserved. The only significant alteration can be identified on the greater trochanter, which is damaged, partially crushed, and eroded. The specimen has a massive body, and its shaft is strong and relatively wide. The nutrient foramen is not visible in the scan. The proximal end contains all its features. The femoral head is complete and strongly curved, whilst the fovea capitis is small and relatively round. Due to the bulging of the head, the neck is distinct and well defined. The greater trochanter is massive but not too bulky, and damaged, as noted above. Nevertheless, this did not much affect its overall height; evidently, it originally reached the level of the head or slightly extended above it. The trochanteric ridge is curved and closes the trochanteric fossa, which is vertical and deep. The lesser trochanter is faint but visible whilst the third trochanter is absent. The caudal side of the shaft shows a rough surface of the linea aspera, extending from the proximal towards the distal end, clearly defined by a lateral ridge and medial ridge. The distal portion of the bone is broken just above the medial and lateral supracondylar tuberosity and missing.

The above-described morphological features are characteristic for the family Suidae (pigs), in particular for the genus Sus [7, 8]. The question remains, which of the two species did it belong: domestic (Sus domesticus) or wild (Sus scrofa)? Both species are anatomically very similar, and often not distinguishable in faunal material [9]. For postcranial elements, morphological criteria are suggestive at best. More reliable metric data require measurable samples, preferably from an adult individual. Both epiphyses at the proximal ends are fused, indicating an adult animal. Although we are limited to a single element, fortunately, measurements of the hindlimbs show less age or sexual dimorphism-related variability than other postcranial elements [10]. Three appropriate measures could be taken (breadth of the proximal end, depth of femoral head, and the smallest diameter of the diaphysis) and compared to wild boar specimens from the reference collection at the Institute for Quaternary Palaeontology and Geology in Zagreb and to metric data for European domestic and wild pigs of the broader region (Radović S, personal data base) [10, 11]. All three parameters (breadth of the proximal end = 75.0 mm, depth of femoral head = 30.7 mm, and smallest diameter of the diaphysis = 29.6 mm) (Fig. 3e) fall within the size range for wild boars.

The direct radiocarbon date of the shaft sample was 4,781 ± 35 years before present, corresponding to the timeframe of the Middle Eneolithic Retz-Gajary culture in the region, based on a series of radiometric dates [12].